The human pre-replication complex is an open complex.

In eukaryotes, DNA replication initiation requires assembly and activation of the minichromosome maintenance (MCM) 2-7 double hexamer (DH) to melt origin DNA strands. However, the mechanism for this initial melting is unknown. Here, we report a 2.59-Å cryo-electron microscopy structure of the human MCM-DH (hMCM-DH), also known as the pre-replication complex. In this structure, the hMCM-DH with a constricted central channel untwists and stretches the DNA strands such that almost a half turn of the bound duplex DNA is distorted with 1 base pair completely separated, generating an initial open structure (IOS) at the hexamer junction. Disturbing the IOS inhibits DH formation and replication initiation. Mapping of hMCM-DH footprints indicates that IOSs are distributed across the genome in large clusters aligning well with initiation zones designed for stochastic origin firing. This work unravels an intrinsic mechanism that couples DH formation with initial DNA melting to license replication initiation in human cells.

CDT1 inhibits CMG helicase in early S phase to separate origin licensing from DNA synthesis.

Human cells license tens of thousands of origins of replication in G1 and then must stop all licensing before DNA synthesis in S phase to prevent re-replication and genome instability that ensue when an origin is licensed on replicated DNA. However, the E3 ubiquitin ligase CRL4Cdt2 only starts to degrade the licensing factor CDT1 after origin firing, raising the question of how cells prevent re-replication before CDT1 is fully degraded. Here, using quantitative microscopy and in-vitro-reconstituted human DNA replication, we show that CDT1 inhibits DNA synthesis during an overlap period when CDT1 is still present after origin firing. CDT1 inhibits DNA synthesis by suppressing CMG helicase at replication forks, and DNA synthesis commences once CDT1 is degraded. Thus, in contrast to the prevailing model that human cells prevent re-replication by strictly separating licensing from firing, licensing and firing overlap, and cells instead separate licensing from DNA synthesis.

Methylation of histone H3 at lysine 37 by Set1 and Set2 prevents spurious DNA replication.

DNA replication initiates at genomic locations known as origins of replication, which, in S. cerevisiae, share a common DNA consensus motif. Despite being virtually nucleosome-free, origins of replication are greatly influenced by the surrounding chromatin state. Here, we show that histone H3 lysine 37 mono-methylation (H3K37me1) is catalyzed by Set1p and Set2p and that it regulates replication origin licensing. H3K37me1 is uniformly distributed throughout most of the genome, but it is scarce at replication origins, where it increases according to the timing of their firing. We find that H3K37me1 hinders Mcm2 interaction with chromatin, maintaining low levels of MCM outside of conventional replication origins. Lack of H3K37me1 results in defective DNA replication from canonical origins while promoting replication events at inefficient and non-canonical sites. Collectively, our results indicate that H3K37me1 ensures correct execution of the DNA replication program by protecting the genome from inappropriate origin licensing and spurious DNA replication.

A dual role for the chromatin reader ORCA/LRWD1 in targeting the origin recognition complex to chromatin.

Eukaryotic cells use chromatin marks to regulate the initiation of DNA replication. The origin recognition complex (ORC)-associated protein ORCA plays a critical role in heterochromatin replication in mammalian cells by recruiting the initiator ORC, but the underlying mechanisms remain unclear. Here, we report crystal and cryo-electron microscopy structures of ORCA in complex with ORC's Orc2 subunit and nucleosomes, establishing that ORCA orchestrates ternary complex assembly by simultaneously recognizing a highly conserved peptide sequence in Orc2, nucleosomal DNA, and repressive histone trimethylation marks through an aromatic cage. Unexpectedly, binding of ORCA to nucleosomes prevents chromatin array compaction in a manner that relies on H4K20 trimethylation, a histone modification critical for heterochromatin replication. We further show that ORCA is necessary and sufficient to specifically recruit ORC into chromatin condensates marked by H4K20 trimethylation, providing a paradigm for studying replication initiation in specific chromatin contexts. Collectively, our findings support a model in which ORCA not only serves as a platform for ORC recruitment to nucleosomes bearing specific histone marks but also helps establish a local chromatin environment conducive to subsequent MCM2-7 loading.

Changing protein-DNA interactions promote ORC binding-site exchange during replication origin licensing.

During origin licensing, the eukaryotic replicative helicase Mcm2-7 forms head-to-head double hexamers to prime origins for bidirectional replication. Recent single-molecule and structural studies revealed that one molecule of the helicase loader ORC (origin recognition complex) can sequentially load two Mcm2-7 hexamers to ensure proper head-to-head helicase alignment. To perform this task, ORC must release from its initial high-affinity DNA-binding site and "flip" to bind a weaker, inverted DNA site. However, the mechanism of this binding-site switch remains unclear. In this study, we used single-molecule Förster resonance energy transfer to study the changing interactions between DNA and ORC or Mcm2-7. We found that the loss of DNA bending that occurs during DNA deposition into the Mcm2-7 central channel increases the rate of ORC dissociation from DNA. Further studies revealed temporally controlled DNA sliding of helicase-loading intermediates and that the first sliding complex includes ORC, Mcm2-7, and Cdt1. We demonstrate that sequential events of DNA unbending, Cdc6 release, and sliding lead to a stepwise decrease in ORC stability on DNA, facilitating ORC dissociation from its strong binding site during site switching. In addition, the controlled sliding we observed provides insight into how ORC accesses secondary DNA-binding sites at different locations relative to the initial binding site. Our study highlights the importance of dynamic protein-DNA interactions in the loading of two oppositely oriented Mcm2-7 helicases to ensure bidirectional DNA replication.

Immunomodulatory potential of cytokine-licensed human bone marrow-derived mesenchymal stromal cells correlates with potency marker expression profile.

Cytokine(s) pre-activation/licensing is an effective way to enhance the immunomodulatory potency of mesenchymal stromal cells (MSCs). Currently, IFN-γ licensing received the most attention in comparison with other cytokines. After licensing human bone marrow-derived MSCs with pro-/anti-inflammatory cytokines IFN-γ, IL-1ß, TNF-α, TGF-ß1 alone or in combination, the in-vitro immunomodulatory potency of these MSCs was studied by incubating with allogeneic T cells and macrophage-like THP-1 cells. In addition, immunomodulation-related molecules filtered by bioinformatics, complement 1 subcomponent (C1s) and interferon-induced GTP-binding protein Mx2 (MX2), were studied to verify whether to reflect the immunomodulatory potency. Herein, we reported that different cytokines cause different effects on the function of MSC. While TGF-ß1 licensing enhances the capacity of MSCs to induce T cells with an immunosuppressive phenotype, IFN-γ-licensing strengthens the inhibitory effect of MSC on T cell proliferation. Both TGF-ß1 and IFN-γ licensing can enhance the effect of MSC on reducing the expression of pro-inflammatory cytokines by M1 macrophage-like THP-1 cells. Interestingly, IFN-γ upregulates potential potency markers extracellular C1s and kynurenine (KYN) and intracellular MX2. These three molecules have the potential to reflect mesenchymal stromal cell immunomodulatory potency. In addition, we reported that there is a synergistic effect of TGF-ß1 and IFN-γ in immunomodulation.

Trends of drug licensing in China: From bring-in to go-global.

Global medicine R&D has shown a stabilized trend after fleeting prosperity lately. Despite of discouraging whole picture, the role of China in cross-border out-licensing activities of medical therapeutics keeps rising, which started since 2020 and further boosted in 2023. A holistic analysis of drug and technology licensing, involving Chinese enterprises from 2019 to 2023, revealed 807 license-in deals and 401 license-out deals. In-licensing showed a decreasing trend, while out-licensing was the opposite. Increasing ingenuity of Chinese products has led to a shift from an import-oriented to an export-oriented situation, such as antibody-drug-conjugates, and cell/gene therapies. With supportive policies of Chinese government, southeast Asia is becoming an important downstream market for out-licensed drugs. Biotech innovators are emerging as a dominant force with core innovation capacity. China is embracing international collaboration by showing favorable markets and high-level innovative products.

Cdt1 overexpression drives colorectal carcinogenesis through origin overlicensing and DNA damage.

Chromatin licensing and DNA replication factor 1 (CDT1), a protein of the pre-replicative complex, is essential for loading the minichromosome maintenance complex (MCM) helicases onto the origins of DNA replication. While several studies have shown that dysregulation of CDT1 expression causes re-replication and DNA damage in cell lines, and CDT1 is highly expressed in several human cancers, whether CDT1 deregulation is sufficient to enhance tumorigenesis in vivo is currently unclear. To delineate its role in vivo, we overexpressed Cdt1 in the mouse colon and induced carcinogenesis using azoxymethane/dextran sodium sulfate (AOM/DSS). Here, we show that mice overexpressing Cdt1 develop a significantly higher number of tumors with increased tumor size, and more severe dysplastic changes (high-grade dysplasia), compared with control mice under the same treatment. These tumors exhibited an increased growth rate, while cells overexpressing Cdt1 loaded greater amounts of Mcm2 onto chromatin, demonstrating origin overlicensing. Adenomas overexpressing Cdt1 showed activation of the DNA damage response (DDR), apoptosis, formation of micronuclei, and chromosome segregation errors, indicating that aberrant expression of Cdt1 results in increased genomic and chromosomal instability in vivo, favoring cancer development. In line with these results, high-level expression of CDT1 in human colorectal cancer tissue specimens and colorectal cancer cell lines correlated significantly with increased origin licensing, activation of the DDR, and microsatellite instability (MSI). © 2022 The Pathological Society of Great Britain and Ireland.

Scope of practice and opioid prescribing behavior of nurse practitioners serving Medicare beneficiaries.

Policymakers aiming to increase access to health care while simultaneously keeping costs low and quality high are considering expanding the practice authority and prescriptive authority of nurse practitioners in order to address primary care shortages. While we know this increases access, some researchers argue that the expansion of job autonomy of nurse practitioners can compromise the quality and safety of rendered medical services. This paper investigates quality and safety outcomes in prescribing behaviors of nurse practitioners who have prescribed opioids for Medicare Part D beneficiaries using a unique source of policy variation, nurse practitioners with the ability to prescribe medication who move to either states with or without physician supervision. We find that scope of practice expansions do not compromise quality and safety in terms of potential abuse or misuse of prescriptive authority.

The effects of dental hygienist autonomy on dental care utilization.

We investigate the effects of regulations governing the practice autonomy of dental hygienists on dental care use with the 2001-2014 Medical Expenditure Panel Survey. We measure the strength of autonomy regulations by extending the Dental Hygiene Professional Practice Index to the years 2001-2014, allowing us to capture changes in regulations within states over time. Using a difference-in-differences framework applied to selected states, we find that relaxing supervision requirements to provide dental hygienists moderate autonomy results in an increase in total dental visits due to greater use of preventive dental care. However, the use of dental treatment decreases when states adopt the highest level of autonomy. Both sets of estimates increase in magnitude when we subset the sample to dental care provider shortage areas. In support of these findings, we show that dental visits shift to dental hygienists in shortage areas when states expand the scope of practice of hygienists, and that there is an increase in tasks performed by hygienists, such as cleanings and dental exams.

Cdt1 Self-associates via the Winged-Helix Domain of the Central Region during the Licensing Reaction, Which Is Inhibited by Geminin.

The initiation of DNA replication is tightly controlled by the licensing system that loads replicative DNA helicases onto replication origins to form pre-replicative complexes (pre-RCs) once per cell cycle. Cdc10-dependent transcript 1 (Cdt1) plays an essential role in the licensing reaction by recruiting mini-chromosome maintenance (MCM) complexes, which are eukaryotic replicative DNA helicases, to their origins via direct protein-protein interactions. Cdt1 interacts with other pre-RC components, the origin recognition complex, and the cell division cycle 6 (Cdc6) protein; however, the molecular mechanism by which Cdt1 functions in the MCM complex loading process has not been fully elucidated. Here, we analyzed the protein-protein interactions of recombinant Cdt1 and observed that Cdt1 self-associates via the central region of the molecule, which is inhibited by the endogenous licensing inhibitor, geminin. Mutation of two ß-strands of the winged-helix domain in the central region of Cdt1 attenuated its self-association but could still interact with other pre-RC components and DNA similarly to wild-type Cdt1. Moreover, the Cdt1 mutant showed decreased licensing activity in Xenopus egg extracts. Together, these results suggest that the self-association of Cdt1 is crucial for licensing.

Permissive beliefs in the context of gaming, online shopping and alcohol drinking - Systematic development of a self-report measure.

BACKGROUND: The activation of permissive beliefs is a cognitive mechanism through which individuals permit themselves to engage in pleasurable, yet potentially unregulated activities. Existing measures are heterogenous, focusing either on specific behaviors or on particular licensing mechanisms. The new Permissive Beliefs Questionnaire (PBQ) seeks to integrate self-licensing mechanisms from various research domains and to be applicable to different behaviors. METHODS: Study 1 aimed at exploring the factor structure and reduce the number of items. In study 2, we used confirmatory factor analysis and tested convergent and discriminant validity in three subsamples of individuals playing videogames (n = 489), shopping online (n = 506), and drinking alcohol (n = 511). We tested the hypothesis whether individuals who experience a self-regulatory conflict show a greater expression of permissive beliefs. RESULTS: The final version of the PBQ consists of 12 items which represent two factors: Deserving Reward and Defensive Optimism. The PBQ exhibited robust model fit indices and internal consistencies in the three samples. Permissive beliefs were heightened among individuals intending to downregulate their gaming, shopping, or drinking behaviors as compared to individuals without this intention. CONCLUSION: The PBQ is a valid measure of permissive beliefs for gaming, online shopping, and drinking alcohol. It serves an ecologic and psychometrically valid tool to address empirical research questions regarding the functioning of permissive beliefs. Additionally, it may be used in clinical settings to measure and raise an understanding for permissive beliefs in clients.

Intellectual property licensing of therapeutics during the COVID-19 crisis: lessons learnt for pandemic preparedness and response.

During the COVID-19 pandemic, intellectual property licensing through bilateral agreements and the Medicines Patent Pool were used to facilitate access to new COVID-19 therapeutics in low- and middle-income countries (LMICs). The lessons learnt from the application of the model to COVID-19 could be relevant for preparedness and response to future pandemics and other health emergencies.The speed at which affordable versions of a new product are available in LMICs is key to the realization of the potential global impact of the product. When initiated early in the research and development life cycle, licensing could facilitate rapid development of generic versions of innovative products in LMICs during a pandemic. The pre-selection of qualified manufacturers, for instance building on the existing network of generic manufacturers engaged during the COVID-19 pandemic, the sharing of know-how and the quick provision of critical inputs such as reference listed drugs (RLDs) could also result in significant time saved. It is important to find a good balance between speed and quality. Necessary quality assurance terms need to be included in licensing agreements, and the potentials of the new World Health Organization Listed Authority mechanism could be explored to promote expedited regulatory reviews and timely access to safe and quality-assured products.The number, capacity, and geographical distribution of licensed companies and the transparency of licensing agreements have implications for the sufficiency of supply, affordability, and supply security. To foster competition and support supply security, licenses should be non-exclusive. There is also a need to put modalities in place to de-risk the development of critical pandemic therapeutics, particularly where generic product development is initiated before the innovator product is proven to be effective and approved. IP licensing and technology transfer can be effective tools to improve the diversification of manufacturing and need to be explored for regional manufacturing for accelerated access at scale in in LMICs and supply security in future pandemics.

Exploring the use of Rasch modelling in "common content" items for multi-site and multi-year assessment.

Rasch modelling is a powerful tool for evaluating item performance, measuring drift in difficulty over time, and comparing students who sat assessments at different times or at different sites. Here, we use data from thirty UK medical schools to describe the benefits of Rasch modelling in quality assurance and the barriers to using it. Sixty "common content" multiple choice items were offered to all UK medical schools in 2016-17, and a further sixty in 2017-18, with five available in both years. Thirty medical schools participated, for sixty total datasets across two sessions, and 14,342 individual sittings. Schools selected items to embed in written assessment near the end of their programmes. We applied Rasch modelling to evaluate unidimensionality, model fit statistics and item quality, horizontal equating to compare performance across schools, and vertical equating to compare item performance across time. Of the sixty sittings, three provided non-unidimensional data, and eight violated goodness of fit measures. Item-level statistics identified potential improvements in item construction and provided quality assurance. Horizontal equating demonstrated large differences in scores across schools, while vertical equating showed item characteristics were stable across sessions. Rasch modelling provides significant advantages in model- and item- level reporting compared to classical approaches. However, the complexity of the analysis and the smaller number of educators familiar with Rasch must be addressed locally for a programme to benefit. Furthermore, due to the comparative novelty of Rasch modelling, there is greater ambiguity on how to proceed when a Rasch model identifies misfitting or problematic data.

How people evaluate individuals who act morally prior to acting immorally: An examination of developmental change in moral evaluation, social preference, and prediction of moral behaviors.

Recent studies have enthusiastically examined the developmental origin of moral self-licensing, which is a tendency to act immorally after acting morally. However, it has not been considered enough how children evaluate personality traits of individuals who show moral licensing behavior and whether there is any developmental change in this evaluation. This study examined the developmental change in moral evaluation, social preference, and prediction of moral behaviors for moral licensing characters as well as moral or immoral characters. In total, 36 5- and 6-year-old children, 36 7- and 8-year-old children, and 58 university students participated in the study. The results revealed that 7- and 8-year-olds and adults evaluated moral licensing characters as more moral and likable than those who behave immorally, unlike 5- and 6-year-olds, who did not distinguish between the immoral and moral licensing characters. Importantly, 7- and 8-year-olds judged the moral licensing character as neutral in both moral evaluation and judgment of social preference, suggesting that they thought the immoral behavior was canceled out owing to prior moral behavior in the moral licensing character. However, adults still judged the moral licensing character as immoral and dislikable. Moreover, children's prediction of moral behavior for all characters showed the same tendency as moral evaluation, whereas adults' prediction was slightly different from their moral evaluation. Taken together, our findings revealed that the evaluation of individuals who show moral licensing behavior changed developmentally, and a moral licensing effect was found when evaluating others' moral traits from around 7 or 8 years of age.

Stores Licensing Scheme in remote Indigenous communities of the Northern Territory, Australia: a meta-evaluation.

A form of food retail regulation called the Stores Licensing Scheme was introduced by the Australian Government in 2007-2022 to ensure food security in remote Indigenous communities of the Northern Territory. We examined evaluations of this Scheme implemented under the Northern Territory National Emergency Response and Stronger Futures Northern Territory Acts. Grey literature search identified nine primary source evaluations. Reported outcomes were extracted and thematic analysis utilised to determine barriers and enablers. Outcomes included improved availability and quality of groceries, financial structures, and retail practices, albeit not consistently reported. Governance and food cost were perceived barriers. Future policy aimed to improve food security through community stores should consider food cost subsidy, measures to incentivise all stores to improve standards, and improved governance arrangements enabling self-determination for Aboriginal and Torres Strait Islander Store Directors.

Does higher performance in a national licensing examination predict better quality of care? A longitudinal observational study of Ethiopian anesthetists.

BACKGROUND: Ethiopia made a national licensing examination (NLE) for associate clinician anesthetists a requirement for entry into the practice workforce. However, there is limited empirical evidence on whether the NLE scores of associate clinicians predict the quality of health care they provide in low-income countries. This study aimed to assess the association between anesthetists' NLE scores and three selected quality of patient care indicators. METHODS: A multicenter longitudinal observational study was conducted between January 8 and February 7, 2023, to collect quality of care (QoC) data on surgical patients attended by anesthetists (n = 56) who had taken the Ethiopian anesthetist NLE since 2019. The three QoC indicators were standards for safe anesthesia practice, critical incidents, and patient satisfaction. The medical records of 991 patients were reviewed to determine the standards for safe anesthesia practice and critical incidents. A total of 400 patients responded to the patient satisfaction survey. Multivariable regressions were employed to determine whether the anesthetist NLE score predicted QoC indicators. RESULTS: The mean percentage of safe anesthesia practice standards met was 69.14%, and the mean satisfaction score was 85.22%. There were 1,120 critical incidents among 911 patients, with three out of five experiencing at least one. After controlling for patient, anesthetist, facility, and clinical care-related confounding variables, the NLE score predicted the occurrence of critical incidents. For every 1% point increase in the total NLE score, the odds of developing one or more critical incidents decreased by 18% (aOR = 0.82; 95% CI = 0.70 = 0.96; p = 0.016). No statistically significant associations existed between the other two QoC indicators and NLE scores. CONCLUSION: The NLE score had an inverse relationship with the occurrence of critical incidents, supporting the validity of the examination in assessing graduates' ability to provide safe and effective care. The lack of an association with the other two QoC indicators requires further investigation. Our findings may help improve education quality and the impact of NLEs in Ethiopia and beyond.

Performance of ChatGPT Across Different Versions in Medical Licensing Examinations Worldwide: Systematic Review and Meta-Analysis.

BACKGROUND: Over the past 2 years, researchers have used various medical licensing examinations to test whether ChatGPT (OpenAI) possesses accurate medical knowledge. The performance of each version of ChatGPT on the medical licensing examination in multiple environments showed remarkable differences. At this stage, there is still a lack of a comprehensive understanding of the variability in ChatGPT's performance on different medical licensing examinations. OBJECTIVE: In this study, we reviewed all studies on ChatGPT performance in medical licensing examinations up to March 2024. This review aims to contribute to the evolving discourse on artificial intelligence (AI) in medical education by providing a comprehensive analysis of the performance of ChatGPT in various environments. The insights gained from this systematic review will guide educators, policymakers, and technical experts to effectively and judiciously use AI in medical education. METHODS: We searched the literature published between January 1, 2022, and March 29, 2024, by searching query strings in Web of Science, PubMed, and Scopus. Two authors screened the literature according to the inclusion and exclusion criteria, extracted data, and independently assessed the quality of the literature concerning Quality Assessment of Diagnostic Accuracy Studies-2. We conducted both qualitative and quantitative analyses. RESULTS: A total of 45 studies on the performance of different versions of ChatGPT in medical licensing examinations were included in this study. GPT-4 achieved an overall accuracy rate of 81% (95% CI 78-84; P<.01), significantly surpassing the 58% (95% CI 53-63; P<.01) accuracy rate of GPT-3.5. GPT-4 passed the medical examinations in 26 of 29 cases, outperforming the average scores of medical students in 13 of 17 cases. Translating the examination questions into English improved GPT-3.5's performance but did not affect GPT-4. GPT-3.5 showed no difference in performance between examinations from English-speaking and non-English-speaking countries (P=.72), but GPT-4 performed better on examinations from English-speaking countries significantly (P=.02). Any type of prompt could significantly improve GPT-3.5's (P=.03) and GPT-4's (P<.01) performance. GPT-3.5 performed better on short-text questions than on long-text questions. The difficulty of the questions affected the performance of GPT-3.5 and GPT-4. In image-based multiple-choice questions (MCQs), ChatGPT's accuracy rate ranges from 13.1% to 100%. ChatGPT performed significantly worse on open-ended questions than on MCQs. CONCLUSIONS: GPT-4 demonstrates considerable potential for future use in medical education. However, due to its insufficient accuracy, inconsistent performance, and the challenges posed by differing medical policies and knowledge across countries, GPT-4 is not yet suitable for use in medical education. TRIAL REGISTRATION: PROSPERO CRD42024506687; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=506687.

Comparing the performance of ChatGPT GPT-4, Bard, and Llama-2 in the Taiwan Psychiatric Licensing Examination and in differential diagnosis with multi-center psychiatrists.

AIM: Large language models (LLMs) have been suggested to play a role in medical education and medical practice. However, the potential of their application in the psychiatric domain has not been well-studied. METHOD: In the first step, we compared the performance of ChatGPT GPT-4, Bard, and Llama-2 in the 2022 Taiwan Psychiatric Licensing Examination conducted in traditional Mandarin. In the second step, we compared the scores of these three LLMs with those of 24 experienced psychiatrists in 10 advanced clinical scenario questions designed for psychiatric differential diagnosis. RESULT: Only GPT-4 passed the 2022 Taiwan Psychiatric Licensing Examination (scoring 69 and ≥ 60 being considered a passing grade), while Bard scored 36 and Llama-2 scored 25. GPT-4 outperformed Bard and Llama-2, especially in the areas of 'Pathophysiology & Epidemiology' (χ2 = 22.4, P < 0.001) and 'Psychopharmacology & Other therapies' (χ2 = 15.8, P < 0.001). In the differential diagnosis, the mean score of the 24 experienced psychiatrists (mean 6.1, standard deviation 1.9) was higher than that of GPT-4 (5), Bard (3), and Llama-2 (1). CONCLUSION: Compared to Bard and Llama-2, GPT-4 demonstrated superior abilities in identifying psychiatric symptoms and making clinical judgments. Besides, GPT-4's ability for differential diagnosis closely approached that of the experienced psychiatrists. GPT-4 revealed a promising potential as a valuable tool in psychiatric practice among the three LLMs.

ChatGPT-4: An assessment of an upgraded artificial intelligence chatbot in the United States Medical Licensing Examination.

PURPOSE: ChatGPT-4 is an upgraded version of an artificial intelligence chatbot. The performance of ChatGPT-4 on the United States Medical Licensing Examination (USMLE) has not been independently characterized. We aimed to assess the performance of ChatGPT-4 at responding to USMLE Step 1, Step 2CK, and Step 3 practice questions. METHOD: Practice multiple-choice questions for the USMLE Step 1, Step 2CK, and Step 3 were compiled. Of 376 available questions, 319 (85%) were analyzed by ChatGPT-4 on March 21st, 2023. Our primary outcome was the performance of ChatGPT-4 for the practice USMLE Step 1, Step 2CK, and Step 3 examinations, measured as the proportion of multiple-choice questions answered correctly. Our secondary outcomes were the mean length of questions and responses provided by ChatGPT-4. RESULTS: ChatGPT-4 responded to 319 text-based multiple-choice questions from USMLE practice test material. ChatGPT-4 answered 82 of 93 (88%) questions correctly on USMLE Step 1, 91 of 106 (86%) on Step 2CK, and 108 of 120 (90%) on Step 3. ChatGPT-4 provided explanations for all questions. ChatGPT-4 spent 30.8 ± 11.8 s on average responding to practice questions for USMLE Step 1, 23.0 ± 9.4 s per question for Step 2CK, and 23.1 ± 8.3 s per question for Step 3. The mean length of practice USMLE multiple-choice questions that were answered correctly and incorrectly by ChatGPT-4 was similar (difference = 17.48 characters, SE = 59.75, 95%CI = [-100.09,135.04], t = 0.29, p = 0.77). The mean length of ChatGPT-4's correct responses to practice questions was significantly shorter than the mean length of incorrect responses (difference = 79.58 characters, SE = 35.42, 95%CI = [9.89,149.28], t = 2.25, p = 0.03). CONCLUSIONS: ChatGPT-4 answered a remarkably high proportion of practice questions correctly for USMLE examinations. ChatGPT-4 performed substantially better at USMLE practice questions than previous models of the same AI chatbot.